Has the Index Fairy Lost Her Powers?

02:35 0 Comments

The Index Valuation Premium
It is accepted wisdom that the addition of a company in a major index leads to a rise in its share price and a subsequently higher valuation of the company on the stock market.

A study of the impact of S&P500 index additions and deletions, between 1990 and 2005, found average abnormal share price returns of +8.8% for additions and -15.1% for deletions¹. Another study focusing on the valuation of premium of S&P500 constituents, found that, in 1997, measured in terms of Tobin’s q ratio, S&P500 firms enjoyed a whopping 40% higher valuation than equivalent companies not in the index². As recently as June 2015, a study by S&P capital IQ, comparing stocks in the Russell2000 index against non-Russell2000 index members found, when comparing the Price/book valuation of companies, a large 62% premium valuation for index members versus non-index members³. It’s as if there’s an Index Fairy whose magic wand has the power to dramatically increase (or decrease) the value of a company.

Why Does the Valuation Premium Matter?
It rather makes a mockery of any concept of ‘efficient markets’ if companies are accorded a significantly higher valuation merely because they are lucky enough to be chosen to be in an index. Inclusion in an index won’t in itself produce higher sales and profits.
The economic distortions caused by the existence of an Index Premium are manifold. It means, for example, that capital is being allocated in favour of certain companies over others for no fundamental reason other than index membership. In terms of the Active Vs Passive investing debate, if companies in an index are overvalued, compared with those not in an index, it represents a compelling argument not to invest in index tracking funds.
What is the Index Premium Today?

Given the economic distortions caused by the existence of the Index Premium, it is of more than academic interest to ask: what the level of the Index Premium is today? In order to quantify this Index Premium, we will focus on comparing the valuation of companies included in the S&P500 index with those excluded.
Table 1: Summary of Analysis of S&P500 Members Vs Non-Members (Meeting the Minimum Capitalisation Requirement)

The raw data, shown in Table 1, compares the characteristics of the S&P500 members against all those (178) companies which have been excluded from the index, despite meeting the S&P Index Committees minimum market capitalisation threshold of $5.3bn. Surprisingly, the raw data suggests, on a number of different valuation measures, an S&P500 index valuation discount rather than premium.
Comparing Apples with Apples
However, in any comparison of valuations, it is important to compare ‘apples with apples’. The S&P500 includes companies with market capitalisations varying between $5bn and $500bn classified between any one of 156 Global Industry Classification (GICS) sub-industries. To measure the index premium more accurately, we have therefore paired companies within a narrow band of capitalisation, i.e. $5.3bn-$10bn, operating in the same industry.
Out of the 134 companies, with market caps between $5.3bn and $10bn, excluded from the index, we were able to identify 27 very close pairings with companies included in the index. For each pairing, we then compared the valuation of both companies against 8 commonly used valuation measures, i.e. Tobin’s q, Price/Book, Consensus Estimated  Price/Earnings (Rolling 12 month estimates), Price/Earnings, Price/Sales, Dividend Yield, Price/Earnings Before Interest Tax Depreciation and Amortization (trailing 12 month) and Price/cash flow. This provided a total of 216 comparative observations.
An Index Discount!!
When we looked at the results it was with a mix of astonishment and disbelief. The results showed not only that there was no Index Premium but there was in fact an Index Discount. Remarkably, all 8 valuation measures show, on average, companies in the S&P500 index are trading on a lower valuation than those excluded from the index. Out of the 216 valuation comparisons, 136 indicated a Discount, 72 a Premium with 8 Equal/NA. Table 2 summarises the findings.

Table 2: Summary of Pairs Analysis

By way of illustration, the median estimated Tobin’s q of stocks in the sample of S&P500 stocks was 1.8 compared with 2.0 for non-members. Weighted by market cap, the comparatives are 2.0 vs 2.2. Similarly, the median P/B is 2.7 for S&P500 members and 2.8 for non-members. Again, weighted by market cap, the comparatives are 3.0 vs 4.0. These figures fly in the face of previous studies which, using both these measures of value, have suggested a large premium valuation for index members versus non-members.
There is no one valuation measure which will give you a definitive answer to what the value of the Index Discount might be. However, on many cap weighted measures (Tobin’s q, PE BF12m, PE, EV/Ebitda) S&P500 included stocks trade on a valuation discount of slightly over 10% to those excluded from the index. Other measures of value indicate a somewhat larger value discount.
Why Has the Index Premium Vanished?
There are a number of possible reasons for the lack of an Index Premium.
  • The most credible reason is that markets are efficient and any Index Premium, if it ever existed, has been arbitraged away by investors. This explanation is consistent with David Blitzer’s (Chairman of the S&P Index’s Committee) observation in July 2015, that there was no ‘pop’ or permanent price increase associated with S&P500 index membership⁵.
  • Beta measures the sensitivity of a securities to moves in the overall market. Compared to the S&P500 market cap weighted beta of 1, the 178 large caps excluded from the index have a lower beta of 0.96.  The recent drop in the S&P500 index will have impacted the share prices of non-index constituents less, reducing any valuation premium (and may even have contributed to the observed discount). The existence of a beta difference, though somewhat smaller than previous studies, is consistent with previous research.⁴
  • Many academic studies have observed a volatility anomaly. Low beta stocks have historically performed better than high beta stocks (counter-intuitive as investors normally imagine a higher risk is associated with higher returns) and  therefore as the beta anomaly is arbitraged away, it may have indirectly boosted the valuation of non-index (lower beta) stocks.
  • Previous studies may have substantially overestimated the Index Premium. The 40% -60% Index Premium suggested in previous research looks excessive and is difficult to reconcile with research into the abnormal returns (or recent lack of them) associated with index additions/deletions. These are, after all, related concepts. In our research, we avoided the temptation of dipping into the pool of mid cap stocks, with capitalisations below the minimum threshold for index inclusion, in order to increase the number of pairings. We also used a broad range of valuation measures. 
How Can We Rationalise an Index Discount?
The Index Discount was a surprising result, but not as inexplicable as it might at first seem. It is possible that the observed Index Discount will prove temporary. The selling of S&P500 index linked Exchange Traded Funds (ETF’s) in January may have pushed the valuation of index constituents below that of non-index constituents. If that is the case, the Index discount may prove temporary and disappear.
The author believes that the lower valuation of index constituents can be explained mainly in terms of subtle ‘value’ and ‘earnings quality’ (predictability and sustainability of earnings) biases in the in the S&P index methodology, specifically:
S&P Requirements: ‘’The sum of the most recent four consecutive quarters’ as-reported earnings should be positive as should the most recent quarter.’’
Comment: Included in the denominator of many measures of value. A clear value screen.
S&P Requirement: ‘’A company’s balance sheet leverage, which should be operationally justifiable in the context of both its industry peers and its business model.’’
Comment: Will exclude companies perceived as financially weak. As at 3.2.2016, S&P500 members had a median debt/equity ratio of 54%, compared with 70% for the 178 large cap non-members, while the Altman z-score (of default risk) was 3.55 against 3.11. These measures confirm the more risky and leveraged nature of non-index stocks.  
S&P Requirement: ‘’Public float of at least 50% of the stock.’’
Comment: May favour companies with longer histories where founders have sold down their holdings.
S&P Requirement: ‘’Contribution to sector balance maintenance.’’
Comment: May favour more established companies already in the index.
S&P Requirement: ‘’Initial public offerings should be seasoned for six to 12 months before being considered for addition to an index.’’
Comment: Again, may favour more established firms.
S&P Requirement: ‘’Ineligible securities include business development companies (BDCs), limited partnerships, master limited partnerships, limited liability companies (LLCs)…’’
Comment: This excludes many hybrid business structures which are sometimes symptomatic of future problems.
S&P Requirement: ‘’… an index constituent that appears to violate criteria for addition to that index is not deleted unless ongoing conditions warrant an index change.’’
Comment: Requires an overall assessment of the financial position of a firm favouring financially stronger firms.
Dividend Policies Reflect the Value/Quality Biases of the Index
Nothing reflects a company’s ‘earnings quality’ better than its dividend policy. In general, companies seek to pay affordable, stable and growing dividends over time. No dividends, or conversely unsustainably high dividend yields, are often indicative of more speculative situations.
One of the most notable results of our research, in tables 1 and 2, is the very low median dividend yield of the non-index companies compared to the S&P500 constituents. While the overall cap weighted yields may look comparable, digging deeper, 40% of the large cap non-index stocks are paying no dividends at all, while 7% (12/178) have dividend yields above 10%. In stark contrast, only 15% of S&P500 constituents are paying no dividend with less than 1% trading on dividend yields above 10%.  Our 27 closely matched pairs tell the same story; with a very low median dividend yield (0.1%) for the non-index companies in our sample compared to the S&P500 constituents (1.8%). Just 14 out of the 27 non-S&P500 companies were paying dividends compared with 22 out of 27 index constituents.
A Cautionary Note
While there would appear to be no Index Premium on the S&P500, and the current Index Discount may have a very rational value/quality explanation, that does not mean that the judgements of the leading Index Committees have no bearing on share prices. During the summer of 2015, there was widespread speculation with regard to the imminent inclusion of mainland Chinese equities in the major MSCI/FTSE World Indices. Significantly, the peak of a powerful rally in these markets coincided with this speculation and the subsequent disappointment at their delayed inclusion (Chart 1). Whilst in general, indices reflect the fundamental performances of their constituents. There are occasions when Index Committees must make important judgements about the underlying suitability of individual titles or markets. In this regard, the Index Fairy hasn’t lost her powers.

Chart 1: Index Committees Can Still Influence Markets

Source: AKRO investiční společnost, a.s./Bloomberg

There is no evidence to suggest an Index Premium. Markets would appear to be efficient with ‘’no money on the table’’ for arbitraging any index anomaly. Investors may therefore invest into the S&P500 index funds knowing that they are not buying an expensive subset of big cap stocks. This conclusion is consistent with recent research suggesting there is no index ‘pop’ and that the performance of index constituents reflects their underlying economic performances rather than any Index Premium. 
Our findings points to an Index Discount, of a magnitude of 10% or more, which probably reflects the value/quality biases in the S&P500 Index methodology. It is plausible that market participants do not appreciate the full extent of these value/quality biases. These differences impact on a broad range of valuation measures, but are most apparent when reviewing the dividend policies, leverage and credit worthiness of index vs non-index constituents.
The research confirmed the higher co-movement of index constituents. Measured by beta, the share prices of companies included in the S&P500 index do move more closely with the index as a whole than equivalent companies excluded from the index.
The author cautions that although there would appear to be no Index Premium in relation to the S&P500, there are situations, e.g. last summer in relation to China A-shares, when index inclusion decisions can have an impact on valuations. 

If interested in the whole paper, please read HERE

Jeremy Monk,
Investment Director, AKRO investiční společnost, a.s.

Disclaimer: This article does not constitute investment advice nor a recommendation to purchase any security.

¹Antti Petajisto. (2008).’ ‘’The index premium and its hidden cost for index funds.’’
²Randall Morck and Fan Yang (2000). ‘’ The Mysterious Growing Value of S&P500 Membership.’’ Tobin’s q ratio compares the market value of a company with the replacement cost of the firm’s assets.
³Norm Alster (Oct. 9, 2015). ‘’The Ease of Index Funds Comes With Risk.’’ New York Times. In this article, the Premium index valuation is ‘’illustrated by the comparison of the freight haulers J.B. Hunt Transport Services and Swift Transportation.’’ It should be noted, the former has a capitalisation of $8.5bn while the latter just $2.4bn. Not only are the companies of a different magnitude of size, the latter company’s capitalisation falls well below the $5.3bn threshold for S&P500 index conclusion.
⁴Jeffrey Wurgler. (2010). ‘’On the Economic Consequences of Index-Linked Investing.’’
⁵David Blitzer (July 1, 2015).’’ Stocks, Not Indices, Move Markets’’. Maria Kasch and Asani Sarkar. (2012). “Is There an S&P 500 Index Effect?” Staff Report No. 484, Federal Reserve Bank of New York. Honghui Chen, Vijay Singal and Robert F. Whitelaw. (2015) “Comovement Revisited”.

0 komentářů: